Biting the AI Apple: Magritte and the AI

Part Three: Text and Image
by Matteo Moriconi in collaboration with Noah Charney

Combining generative art using Midjourney with ChatGPT, given the prompt of "The beauty of 'Ceci n'est pas une pomme" when attributed to Walt Disney produced a work that is Magrittean in its surrealism. The software recognized that a statement that begins with "Ceci n'est pas une" in the context of art should have a Magrittean vibe. That is some serious semantic understanding for a machine to grasp what the user was after.

Manga and Surrealism share similarities in their approach to art and storytelling. Both are known for their bold and imaginative imagery, which often blurs the lines between reality and fantasy. One of the key similarities is their use of symbolism-this relates to the semiotics mentioned earlier. Both manga and Surrealist art rely heavily on symbols and motifs to convey deeper meanings and messages. These symbols often have multiple layers of interpretation and are used to explore complex themes such as identity, sexuality and social issues. Both manga and Surrealism also challenge the traditional conventions of storytelling and representation. Manga often uses non-linear and unconventional narrative structures, while Surrealism is known for its use of dream-like imagery and the exploration of the subconscious. Another parallel between manga and Surrealism is their ability to create a sense of disorientation and unease in the viewer. Both use bizarre and unsettling imagery to create a sense of unease and to challenge the viewer's understanding.

In the age of artificial intelligence, surrealism might be redefined as the study and portrayal of the strange and unexpected using technology and AI. This can involve using artificial intelligence to produce surrealistic art or writing or to study and build upon established surrealism techniques and topics. In addition, the idea of the "uncanny valley" in artificial intelligence research, where a robotic or digital entity appears nearly human but falls short, could be viewed as a type of surrealism in the age of artificial intelligence.
Surrealism fits squarely into the real estate of the "uncanny valley." The uncanny valley is a concept in robotics and computer graphics that describes a phenomenon in which, as robots or avatars (or AI software) become more human-like in appearance and behavior, a point is reached at which they become, for lack of a more elegant term, creepy, from a human perspective.
The term was first coined by Masahiro Mori in 1970, a scientist working on robots. It refers to the dip in a graph that shows the emotional response of humans to artificial entities as they become more human-like. In theory robots could become so human-like as to be unrecognizable as robots. We're already there when it comes to some responses from software, when you can't really tell if you're chatting with a person or a chatbot. But there is a threshold at which the robot or AI software is still recognizable as non-human, but human enough to be disquieting.
Has AI already reached this point?

To answer this question, I asked ChatGPT to write a letter from Surrealist printmaker M. C. Escher to Rene Magritte on the subject of an artwork based on Escher's Drawing Hands and a Magrittean apple.

Dear Rene,
I have depicted two hands drawing an apple, with the apple itself being a representation of the famous painting of yours, The Son of Man. I was inspired by the surreal imagery and the play on perception in your painting, and the idea of hands drawing themselves and the object they are drawing. The theme of self-reference and creation is one that I believe resonates strongly in both of our works.
Sincerely,
M.C. Escher

Detail: Drawing Hands by M. C. Escher, 1948, Lithograph.

Generative art of the future will combine self-reference and creativity, two factors that indeed resonate in the work of Escher and Magritte. Magritte's body of work is rich in meaning and ideas that are conveyed through the use of semiotics, defamiliarization, concealment, and the uncanny. His paintings challenge our perceptions of reality and encourage us to question the nature of representation, language, and identity. The same could be said for AI software, as it creates art and text that inhabits this uncanny valley, drifting into material that is unrecognizable as artificial, surreal for its ability to instill awe in us, marveling at the capability of a machine to magic up such wonders.